Barry T. (Ty) Wilson

• Geospatial modeling and mapping using forest inventory data
• Small area estimation using a model-based mode of inference
• Nonparametric models, such as the k-nearest neighbors estimator
• Hierarchical Bayesian models and MCMC methods
• Use of dense time series of satellite imagery (e.g. MODIS & Landsat) to track land cover change and forest disturbance

My line of research focuses on enhancing the value of the foundational information provided by the FIA program. Rather than requiring our users to become familiar with our complex database of inventory data, we can provide geospatial raster products as outputs from our modeling efforts. These data can be used by scientists, decision makers, and the interested public for multiple purposes, ranging from the preparation of state forest assessments and wildlife habitat models, to maps of current tree species distributions and the projection of future forest conditions. Another focus of my research is on characterizing the uncertainty of these raster products, providing the basis for making statistical inferences about forests from our geospatial models.